Realization and Optimization of Face Detection Algorithm Based on DSP

It introduced the principle and the process of human face detection algorithm, as well as its implementation and optimization using DSP. It applied AdaBoost algorithm to detect human face. Before the detection, some methods were used to optimize original image data, such as image scale, median filter, histogram equalization and edge detection. After the detection, it judged and modified the results to improve the detection accuracy further. Then it used many optimization methods to meet the real-time requirement. Experimental results show that, this DSP system can meet real-time requirement even if it detects faces in 720×576 high-resolution images, at the same time it has a high accuracy and low miss rate in a complex environment.